Be specific or be invisible
AI systems form category associations from repetition
Large language models learn associations during training and reinforce them at inference time through retrieval. If a single brand is mentioned, repeatedly and consistently, alongside a specific phrase — 'CRM for HVAC contractors', 'shoulder physiotherapy in Birmingham', 'open-source vector database' — the model develops a strong link between brand and category. When asked about that category, it produces the brand. The opposite is also true: a generic brand that calls itself 'a marketing partner for ambitious businesses' is associated with nothing in particular and is rarely the first name returned.
Specificity is a forcing function
Specific positioning — 'we help SaaS companies between one and ten million pounds in annual revenue scale through performance marketing' — does three things at once. It tells AI systems exactly which queries to surface the brand for. It tells potential customers immediately whether they are in the target market. And it forces internal discipline: the brand must actually be best for that segment, because the moment it tries to serve everyone again, the association weakens. The cost is real — turning away clients who do not fit — and that cost is precisely why most competitors will not pay it.
Comparison and category guides build defaults
Pages that explicitly compare options inside a category — 'the five best CRMs for HVAC contractors', 'how to choose a plumbing-firm bookkeeping system' — are heavily cited by AI assistants for evaluation queries. They serve a real reader need and they create a corpus of content that AI systems associate with the brand. Honest comparisons that include the brand alongside competitors usually outperform self-promotional listicles, because AI systems weigh independence and tend to discount obviously biased pages. The discipline is to compete on craft: write the comparison your customer actually needed, even when the conclusion sometimes favours another product. The keyword tools directory lists tools that surface the specific category queries worth writing about.
The flywheel rewards category leaders disproportionately
Once a brand becomes the default recommendation, three reinforcing effects compound. More agent traffic produces more user interactions, which produces more data, which improves products and content. More citations on third-party sites — reviews, comparisons, news — strengthen the AI association. More direct queries by name reinforce the brand-category link in retrieval. The brands that win this flywheel in early 2026 to 2027 are likely to remain default recommendations for years, because each subsequent challenger has to dislodge an entrenched association rather than compete on a level field. The window covered in lesson 15 is, in large part, the window to enter this flywheel.
Things to consider — B2B and B2C build authority differently
The flywheel works in both consumer and business markets, but the inputs differ. B2C category authority is fed by consumer reviews, social mentions, comparison content, and the steady drip of unboxing and how-to posts on third-party platforms. The buying cycle is short, the decision is often made by one person, and AI assistants pick up the brand quickly once consumer signal accumulates. B2B authority is fed by white papers, analyst placements, podcast guest appearances, technical comparisons by independent practitioners, and citations in trade press. The buying cycle is months, the decision is multi-stakeholder, and AI assistants build the association more slowly because the public signal is thinner per buyer. The implication is operational: a B2C brand investing only in white papers builds the wrong kind of authority, and a B2B brand chasing consumer reviews wastes effort. Match the inputs to the market.
Imagine an AI assistant asked, 'who is the best at X?' for the most specific version of what your business does. What three brands would it name? How did it decide?